The Jobs-To-Be-Done (JTBD) Framework: A Practical Guide
By Krishna Vepakomma
Sales & AI Expert
By Krishna Vepakomma
Sales & AI Expert

People do not buy products. They "hire" them to make progress in a situation they are stuck in. That is the core idea behind Jobs-To-Be-Done (JTBD): the unit of analysis is not the customer's demographics or the product's features, but the job the customer is trying to get done. Understand the job well enough and your product decisions, messaging, and sales conversations all get sharper. This guide covers what a job is, how to write a job statement, how to interview for one, and how to actually use it.
A job has three dimensions, and skipping any of them leads to shallow conclusions:
A spreadsheet does the functional job. It fails the emotional and social ones the moment a deal slips through the cracks. That gap — functional job met, emotional and social jobs unmet — is where better products win.
Jobs are also stable while solutions churn. The job "get a meal I do not have to cook" has been hired out to diners, frozen dinners, and delivery apps across decades. Anchoring on the job instead of the current solution is what keeps you from being disrupted by the next one.
JTBD models a purchase as a struggle between four forces:
A switch happens only when push plus pull beats anxiety plus habit. This is a genuinely useful lens for sales and marketing, because it tells you the two levers you often ignore: you can win not only by increasing pull (more features, better pitch) but by reducing anxiety (proof, guarantees, easy onboarding) and weakening habit (making migration trivial).
A clean job statement follows the form: verb + object + context, phrased in the customer's terms and free of any solution.
Notice the structure: a triggering situation ("when…"), a motivation ("I want to…"), and an intended outcome ("so I…"). If you can only describe the job by naming your product, you have not found the job yet.
You do not discover jobs from surveys or feature requests. You discover them by reconstructing the timeline of a real, recent purchase. The most useful interview walks a customer backward from the moment they started using something new:
The specifics matter more than the generalities. "I lost a deal because two of us emailed the same lead" is a job. "I wanted to be more efficient" is noise.
A five-person startup sells a tool and captures leads from a website form, WhatsApp, and Facebook Lead Ads. Interviews with three churned trial users reveal the same story: leads landed in three inboxes, two reps sometimes contacted the same person, and follow-ups slipped.
Turn that into a job statement: "When leads arrive from several channels at once, I want them unified and assigned automatically, so no lead is contacted twice or forgotten."
Now apply the forces:
This reframes the roadmap. The winning move is not another dashboard (more pull) — it is proving complete multi-channel capture (kill the anxiety) and making setup take minutes (beat the habit). The same statement rewrites the sales pitch: lead with the duplicate-contact pain, not a feature list. One well-built job statement drove product, positioning, and sales at once.
A few traps catch teams the first time they try this:
In B2B, one purchase usually serves several jobs at once. The end user's job might be "stop losing leads," while the manager's job is "get visibility into the team's pipeline," and the approver's job is "not overspend on tools that go unused." These jobs can conflict — the feature that delights the user may be the one the approver sees as risk. Mapping the distinct job for each person in the buying committee, and the forces acting on each, tells you which objection to address for whom. A pitch that nails the user's job but ignores the approver's anxiety stalls at exactly the moment budget gets approved.
A job is only useful if it changes what you do:
JTBD works best when you can see the real situations that trigger a switch, and most teams cannot, because the evidence is scattered across channels and disconnected from what people do next. Inleads pulls the raw material for job discovery into one place. Multi-channel capture (web forms, WhatsApp, Facebook Lead Ads, LinkedIn, NPS, API/SDK) collects the touchpoints where a job first shows up, and the customer data platform unifies them into a single profile, so you can see the actual sequence a customer took rather than a fragmented guess. NPS capture is a direct line to the emotional and social sides of a job — why someone would or would not recommend you. Because product analytics and the CRM live together, you can tie what a person said about their job to what they then did in the product, using the AAARRR funnel to spot exactly where an unmet job causes people to stall. The AI copilot built into the CRM can help synthesize notes and NPS responses into recurring themes. See how capture works on the lead capture page and the analytics on the product analytics page.
JTBD is the idea that customers "hire" a product to make progress in a specific situation. Instead of analyzing who the customer is or what features they ask for, you study the job they are trying to get done — functional, emotional, and social — and build and sell around completing it.
Personas describe who the customer is (demographics, goals, traits). JTBD describes what the customer is trying to accomplish and why they switch solutions. The two complement each other: personas tell you who to talk to, and jobs tell you what will make them act. Many teams use both.
Use the form "verb + object + context" in the customer's own words, with no mention of your product — for example, "when leads arrive from several channels, I want them in one place, so I never contact someone twice." If you can only describe the job by naming your product, you have not found the real job yet.
Push of the current situation, pull of the new solution, anxiety about the new solution, and habit of the present. A customer switches only when push plus pull outweighs anxiety plus habit, which means you can win by reducing anxiety and habit, not only by adding pull.
Walk a customer backward through a real, recent purchase: when they first felt the need, what frustrated the old solution, the specific event that made them look, what almost stopped them, and what "working" looked like afterward. Concrete incidents reveal the job; vague goals like "be more efficient" do not.
Discover more insights about AI sales tools and analytics

